hmm. Now I smoked the stereo amp. There is something unusual happening in the sync circuit between the two amps.
I happen to have a bunch of those 2.1 amps sitting around here that I didn't use because of the crossover network. I like Graham's idea of bypassing the crossover network. Basically I would just do the same input mods that I did to these two to the chips on the 2.1 board?
This is turning into much more of a project than I have time for
I happen to have a bunch of those 2.1 amps sitting around here that I didn't use because of the crossover network. I like Graham's idea of bypassing the crossover network. Basically I would just do the same input mods that I did to these two to the chips on the 2.1 board?
This is turning into much more of a project than I have time for
good idea. are there any unusual behaviors for the sync line? I've seen some posts about errors in the datasheet. The strange thing is that everything was working before.....
Sometimes just the mechanical manipulation of boards around makes shorts... ugh. 9 volts it is!
I did make one change, the SDZ pin was previously being triggered with 5v to turn the amps on without a thump. I changed it to being triggered with 24v (simple on/off switch, while leaving the actual amp power constant)
Back to board modding. I have an extra of these boards that was a pretty nice AM radio. I'm sure the amp is fine, the issue was probably in the opamp which I removed anyway.
Off to investigate.
I will put the mono amp back to master and check for the interference first.
Sometimes just the mechanical manipulation of boards around makes shorts... ugh. 9 volts it is!
I did make one change, the SDZ pin was previously being triggered with 5v to turn the amps on without a thump. I changed it to being triggered with 24v (simple on/off switch, while leaving the actual amp power constant)
Back to board modding. I have an extra of these boards that was a pretty nice AM radio. I'm sure the amp is fine, the issue was probably in the opamp which I removed anyway.
Off to investigate.
I will put the mono amp back to master and check for the interference first.
The datasheet says about the SDZ-pin "Shutdown logic input for audio amp (LOW = outputs Hi-Z, HIGH = outputs enabled). TTL logic levels with compliance to AVCC." AVCC is up to 26V. So no, it is hardly the SDZ-pin that caused the smoke.
If your synchronization fails, the carrier switching may be so slow that your output filter chokes saturate and your chip output is very heavy loaded. But what I understand from what you write, your stereo board was set up to be the master. And, it was the stereo board that burned.
In case of any doubt about what voltage a logical input will accept, put a 47K-100K resistor in series to limit eventual fault current. That may save the IC.
Check your connections well before putting power on and start testing amplifiers with 9-12V
If your synchronization fails, the carrier switching may be so slow that your output filter chokes saturate and your chip output is very heavy loaded. But what I understand from what you write, your stereo board was set up to be the master. And, it was the stereo board that burned.
In case of any doubt about what voltage a logical input will accept, put a 47K-100K resistor in series to limit eventual fault current. That may save the IC.
Check your connections well before putting power on and start testing amplifiers with 9-12V
Last edited:
Yes, i was just thinking about a current limiting resistor. I think there was nothing between. The sync circuit seems a bit confusing to me in the datasheet. is there a good description of it anywhere?
Incidentally i put the mono amp back to being master, tied the SDZ high and removed the burned board.
A test with 12v and the backplane board yielded much less hiss, but the interference is still there. Not constant like it was before, but dependent on the position of the wifi antenna on the source board.
I made a movie of the sound, but I don’t see any way to share the file. How do people do that?
Incidentally i put the mono amp back to being master, tied the SDZ high and removed the burned board.
A test with 12v and the backplane board yielded much less hiss, but the interference is still there. Not constant like it was before, but dependent on the position of the wifi antenna on the source board.
I made a movie of the sound, but I don’t see any way to share the file. How do people do that?
Attachments
Please see Figure 37 in the datasheet https://www.ti.com/lit/ds/symlink/tpa3116d2.pdf?ts=1610848555929&ref_url=https%253A%252F%252Fwww.google.com%252F for how to synchronize two TPA3116D2 chips. Master/Slave set by resistor combinations.
Your new ground-plane will hardly help on hiss. Hum, yes. With hiss, from the MiniDSP or to high TPA3116 gain?
Antennas are made for radiating noise! Shield the electronics from antenna radiation.
How to post videos and sound files I do not know.
Bedtime in Europe.
Your new ground-plane will hardly help on hiss. Hum, yes. With hiss, from the MiniDSP or to high TPA3116 gain?
Antennas are made for radiating noise! Shield the electronics from antenna radiation.
How to post videos and sound files I do not know.
Bedtime in Europe.
Last edited:
Brian,
If you try to synchronize two (or more) TPA3116 chips, be aware that some boards erroneously ground the sync-pin in which case you need to lift the pin free of the board.
For a start I would leave all as "Master". There are very good TPA3116 boards, like the XH-M590, that include two chips both running as master. With the far more important performance issues you have I would not worry about the sync detail for the moment.
If you try to synchronize two (or more) TPA3116 chips, be aware that some boards erroneously ground the sync-pin in which case you need to lift the pin free of the board.
For a start I would leave all as "Master". There are very good TPA3116 boards, like the XH-M590, that include two chips both running as master. With the far more important performance issues you have I would not worry about the sync detail for the moment.
Last edited:
Good news from my side:
Tested the new modded versions of both boards, no smoke (this should not be exciting but I was starting to question my sanity)
I did indeed forget to turn the gain down from 36db on the amps I was hearing the hiss.
When I tested the subwoofer channel using the sub itself, no interference was audible at all. Probably the frequency response of the sub not being sensitive enough at those frequencies.
Next, I fix the gains, and work on the sync.
I was doing some strategic disconnecting of the signal wires to determine the source of the (non hum, non hiss) interference. The Source board (Arylic) with both bluetooth and wifi is the cause. The board does not need to be connected, but only powered on.
I have a few ideas, maybe folks can chime in:
1: install some EMI shielding around the radio board and antennas. I have external antennas anyway and the board is in a convenient place to isolate (the enclosure is metal). Moving the Wifi antenna around appears to have an impact.
2: isolate the board electronically? I have seen in many layouts where the designers divided up the ground plane between the audio circuits and the rest of the design by routing either physical slots or just breaks in the copper with 0 ohm resistors to join them (in fact, this is done in many places on the Arylic board, so I think EMI is a problem there) I think this is probably already being done by the fact that there is now a single point of ground between the arylic board and my 'ground plane', so it may not benefit
3: Switch to an isolated power supply Graham suggested switching to an isolated power supply. I use these routinely for sensitive detector amplifiers. I have a couple that I could rig up for a test. However, I have powered this board from an external supply (separate from the main supply) and not noticed a difference in the interference.
Finally, I uploaded a short video of the interference to youtube here: Interference - YouTube The ground loop hum is one issue, but listen for the "ppppt ppppt", That's the killer. I've been able (before all the smoked boards) to get rid of all of the interference except for that one.
I couldn't get my cheapie "Siglent" scope configured correctly to look at the 24v power signal with any sensitivity so I don't know if that is coming from the SMPS. The interference is present with 2 different SMPS that I have tried, both at 12v and 24v. I have the Tektronics in the other room here, but don't want to bother with it if it isn't a power supply problem.
Tested the new modded versions of both boards, no smoke (this should not be exciting but I was starting to question my sanity)
I did indeed forget to turn the gain down from 36db on the amps I was hearing the hiss.
When I tested the subwoofer channel using the sub itself, no interference was audible at all. Probably the frequency response of the sub not being sensitive enough at those frequencies.
Next, I fix the gains, and work on the sync.
I was doing some strategic disconnecting of the signal wires to determine the source of the (non hum, non hiss) interference. The Source board (Arylic) with both bluetooth and wifi is the cause. The board does not need to be connected, but only powered on.
I have a few ideas, maybe folks can chime in:
1: install some EMI shielding around the radio board and antennas. I have external antennas anyway and the board is in a convenient place to isolate (the enclosure is metal). Moving the Wifi antenna around appears to have an impact.
2: isolate the board electronically? I have seen in many layouts where the designers divided up the ground plane between the audio circuits and the rest of the design by routing either physical slots or just breaks in the copper with 0 ohm resistors to join them (in fact, this is done in many places on the Arylic board, so I think EMI is a problem there) I think this is probably already being done by the fact that there is now a single point of ground between the arylic board and my 'ground plane', so it may not benefit
3: Switch to an isolated power supply Graham suggested switching to an isolated power supply. I use these routinely for sensitive detector amplifiers. I have a couple that I could rig up for a test. However, I have powered this board from an external supply (separate from the main supply) and not noticed a difference in the interference.
Finally, I uploaded a short video of the interference to youtube here: Interference - YouTube The ground loop hum is one issue, but listen for the "ppppt ppppt", That's the killer. I've been able (before all the smoked boards) to get rid of all of the interference except for that one.
I couldn't get my cheapie "Siglent" scope configured correctly to look at the 24v power signal with any sensitivity so I don't know if that is coming from the SMPS. The interference is present with 2 different SMPS that I have tried, both at 12v and 24v. I have the Tektronics in the other room here, but don't want to bother with it if it isn't a power supply problem.
Ah, OK. Made some significant progress!
It ends up that I must have left the ground connected when I tested powering the Arylic board with an external supply. Isolating the arylic (source) board was the ticket.
I wonder if the 5v buck converter supply was causing some issues. I will go and get one of the isolated dc dc converters to solve that. (Thanks Graham!). Unfortunately, all of mine are only 12v-> +5,-5.
Is there a simple way to isolate the ground and power without buying another converter? (shipping time/cost for such a cheap thing kills me)
Now to kill the remaining Hiss from the stereo amp. Even with only 26db gain, (and 4.7k resistors to ground) I am getting some hiss. still an input impedance mismatch?
I am running 26db gain, with a .47uF decoupling capacitor in differential mode.
It ends up that I must have left the ground connected when I tested powering the Arylic board with an external supply. Isolating the arylic (source) board was the ticket.
I wonder if the 5v buck converter supply was causing some issues. I will go and get one of the isolated dc dc converters to solve that. (Thanks Graham!). Unfortunately, all of mine are only 12v-> +5,-5.
Is there a simple way to isolate the ground and power without buying another converter? (shipping time/cost for such a cheap thing kills me)
Now to kill the remaining Hiss from the stereo amp. Even with only 26db gain, (and 4.7k resistors to ground) I am getting some hiss. still an input impedance mismatch?
I am running 26db gain, with a .47uF decoupling capacitor in differential mode.
Good progress.
EMI shielding of noisy circuits, like the Ayrillic, with metal sheet is the traditional way to counter radiation problems. Yes, it is irritating noise. Non-isolated traditional Buck converters can cause a lot of noise.
Impedance mismatch is normally something you relate to higher frequencies. In the audio-band it is often a matter of keeping the impedance-level as low as possible (makes a circuit less noise sensitive) but still high enough for the signal to pass with the amplitude needed.
EMI shielding of noisy circuits, like the Ayrillic, with metal sheet is the traditional way to counter radiation problems. Yes, it is irritating noise. Non-isolated traditional Buck converters can cause a lot of noise.
Impedance mismatch is normally something you relate to higher frequencies. In the audio-band it is often a matter of keeping the impedance-level as low as possible (makes a circuit less noise sensitive) but still high enough for the signal to pass with the amplitude needed.
Strangely enough, it appears to be the cable going from the Arylic to the DSP that is acting as an antenna. The interference is the same if the Arylic is connected or not. But it does have to be turned on.
I wonder if a shielded cable (now using twisted pair)would be easier than an EMI shield. Both would need to be grounded. But to which ground? My first thought would be the isolated ground for the arylic itself. What do you think?
Update: setup beautifully re-worked, isolated (buck) power supply for only the arylic installed: still the odd pulsing interference. It still goes away if powered by USB. Quality of switching power supply doesnt appear to matter. Tried 3 different ones.
Here’s a question: i wonder if there is some interaction between the SMPS and the buck converter (albeit isolated) that only manifests through the radio? All of the data seems to support some interaction between the supplies.
I’ll compare the output from the 5v rail in my setup to the 5v usb supply i have tomorrow with the scope. I’m wondering if there is some additional filtering that i can do between the SMPS and the buck converter (or perhaps after).
Maybe i find a different topology of 5v supply (LDO?)
What do you suggest?
Here’s a question: i wonder if there is some interaction between the SMPS and the buck converter (albeit isolated) that only manifests through the radio? All of the data seems to support some interaction between the supplies.
I’ll compare the output from the 5v rail in my setup to the 5v usb supply i have tomorrow with the scope. I’m wondering if there is some additional filtering that i can do between the SMPS and the buck converter (or perhaps after).
Maybe i find a different topology of 5v supply (LDO?)
What do you suggest?
OK, My hypothesis was correct. I was reading up a bit on buck converters and apparently they can create noise on their supply lines.
I looked at the scope traces with the buck converter (12v->5v) inline and the same ugly ~600khz peaks are present in the upstream SMPS rail and in the downstream Buck regulated 5v and also in the audio signal going to the DSP.
When removing the buck converter and powering the arylic with an external usb supply, the peaks are very small and only present in the incoming SMPS rail.
I think that there must be some ringing or constructive interference going on between the two switching supplies that amplifies the peaks (not sure the EE term).
Here is the Buck converter I have been using: Blocked
The switching frequency is 580kHz. big surprise there.
Now I need to figure out what filtering to use to reduce/eliminate the interference. but at least now I have something to look for.
I looked at the scope traces with the buck converter (12v->5v) inline and the same ugly ~600khz peaks are present in the upstream SMPS rail and in the downstream Buck regulated 5v and also in the audio signal going to the DSP.
When removing the buck converter and powering the arylic with an external usb supply, the peaks are very small and only present in the incoming SMPS rail.
I think that there must be some ringing or constructive interference going on between the two switching supplies that amplifies the peaks (not sure the EE term).
Here is the Buck converter I have been using: Blocked
The switching frequency is 580kHz. big surprise there.
Now I need to figure out what filtering to use to reduce/eliminate the interference. but at least now I have something to look for.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.